The computational simulation of tsinperatursprogrammed gas Chromatographic behavlor of two serles of analytes Is Investigated. The approach presented accounts for both retention-time and peak-width behavlor over diverse temperature programs. The effect of the temperature dependence of carrier gas vlscoslty Is Included In the treatment. Sbnulatlon accuracies are generally 1 % for retention and 10-15 % for peak widths. The isothermal retention t h e data used as input of the simuiatlon are most conveniently reduced to two thennodynamlc quantltles AS, and AHv whkh are proposed as a new class of retention Indexes.Simulation of gas chromatographic behavior has been reported under three general approaches. Classical or semiclassical theoretical treatments (1-5) describe the dependence of separation efficiency of unspecified analytes on changes in column length (6, 7) and cross-section (8,9), carrier pressure (IO), and instrumental inefficiency (11). These treatments do not attempt to simulate particular separations but result in requirements on those that do. A second approach adjusts isothermal retention times (12)(13)(14) or retention indexes (15-23) for oven temperature changes, generally using semiempirical or linear extrapolative methods which predict retention behavior over a limited range of available conditions. The present work follows the third approach, the simulation, whether thermodynamic or empirical, of physical and chemical processes occurring in the instrument (24-31). This approach, while computationally demanding, should afford simulation of specific separation processes and their results over a very wide range of conditions. The degree of success of properly applied, thermodynamically based direct simulation reflects the accuracy of current chromatographic theory.In this work, isothermal data were used to perform temperature-programmed experiments off-line, that is, on a computer rather than on a chromatograph. The accuracy of this method for two classes of analytes is demonstrated. The degree of reproducibility allows relatively small effects such as carrier-gas-viscosity temperature dependence to be verified and included in the treatment.
EXPERIMENTAL SECTIONApparatus. Chromatograms were obtained on two Hewlett-Packard 5890A gas chromatographs with electron-capture detection. Temperatures from the chromatographs' controllers were used directly. No corrections for heating lag time (28) were applied to this capillary system. Organochlorine pesticides (OCPs) were eluted from a 30-m, 0.32 mm i.d. DB-17 capillary column (J&W Scientific) of f i l m thickness 0.25 wm. Nitroaromatic (NA) analytes were eluted from a 30-m, 0.53 mm i.d. DB-210+ column (J&W) of film thickness 1.0 pm. Helium was used as carrier for all experiments.Procedures. Peak retention times were measured at the detector signal maximum as indicated by the Hewlett-Packard 3392A integrator. One-day retention time reproducibility was to 0.01 YO, multiday reproducibilities were to 0.05%, and multiday reproducibilities of retention time ratios...